A computerized cognitive test battery was undertaken (CogState™, Melbourne, Victoria, Australia), which has previously been described in detail [6, 8] and validated in HIV-infected subjects [9]. In brief, all tasks within the battery were adaptations of standard neuropsychological and experimental psychological tests, which assessed a range of cognitive functions. This battery assessed the following domains: detection, identification, monitoring and matched learning (all assessed via speed of test); associate learning and working memory (assessed Idelalisib concentration via accuracy of test); and executive function (assessed via number of errors made

on testing). The battery consisted of tasks in the form of card games. Therefore, subjects

needed only to have an understanding of playing cards, thereby minimizing language and cultural differences among study subjects. Card game instructions were translated into the local language. All study participants HSP cancer completed one full practice test prior to undertaking the study examination to obtain optimal performance at baseline [10]. Statistical analyses were conducted with sas version 9.13 (SAS, Cary, NC) and stata version 10.1 (Statacorp, College Station, TX) and analysis was conducted according to CogState™ recommendations. Reaction times were log10-transformed because of a positive skew of the distribution, and accuracy measures were transformed using arcsine-root transformation. Change scores were calculated for each subject, and these scores standardized according to the within-subject standard deviation (SD). Changes in performance for arms 2 and 3 compared with arm 1 were standardized with a pooled SD, and this was used as the outcome variable in linear regression models to calculate an overall check details effect size for the difference between treatment groups. Composite scores were calculated overall and for the speed and accuracy domains based on the average of standardized scores, and composite changes from baseline scores to weeks 24 and 48 were calculated based on the average of standardized reaction time and accuracy scores. Of 30 subjects enrolled in the

study, 28 completed NC testing at baseline, week 24 and week 48 and were included in this analysis (nine, eight and 11 subjects in arms 1, 2 and 3, respectively). Two subjects who completed baseline NC testing did not attend for follow-up study visits. CD4 lymphocyte count (SD) rose over the 48-week study period from 218 (87) to 342 (145) cells/μL at baseline and week 48, respectively. Other baseline characteristics have previously been described [6]. Of interest, all subjects apart from one had undetectable plasma HIV RNA (<50 HIV-1 RNA copies/mL) at week 48. All statistical results described are unchanged when adjusted for the one subject with detectable plasma HIV RNA at week 48. Overall, improvements in NC function were observed by week 24 and continued to week 48 (Table 1).